VBIM-Net: Variational Born Iterative Network for Inverse Scattering Problems

2024-05-29

arXiv.org

AI Summary

To address low reconstruction accuracy, poor generalization, and weak noise robustness in full-wave inverse scattering, this paper proposes the Variational Born Iterative Network (VBIM-Net). VBIM-Net deeply embeds the physics-based variational Born iterative model into a learnable architecture, enabling alternating updates of the total electric field and dielectric contrast across multiple subnetworks, with explicit incorporation of analytical variational operators. A U-Net–enhanced residual modeling framework eliminates strict alignment requirements between measurement dimensions and grid resolution. Innovatively, soft physics-constrained supervision and joint noise-robust training are introduced, yielding a multi-level composite loss. Evaluated on both synthetic and experimental datasets, VBIM-Net achieves significant improvements in dielectric parameter reconstruction fidelity, generalization capability, and noise resilience—delivering high-accuracy electromagnetic imaging that is physically interpretable, data-efficient, and deployment-flexible.

Technology Category

• Computer Vision: Learning & Optimization for CV
• Machine Learning: Bio-inspired Learning
• Search and Optimization: Mixed Discrete/Continuous Search

Application Category

• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Recently, studies have shown the potential of integrating field-type iterative methods with deep learning (DL) techniques in solving inverse scattering problems (ISPs). In this article, we propose a novel Variational Born Iterative Network, namely, VBIM-Net, to solve the full-wave ISPs with significantly improved structural rationality and inversion quality. The proposed VBIM-Net emulates the alternating updates of the total electric field and the contrast in the variational Born iterative method (VBIM) by multiple layers of subnetworks. We embed the analytical calculation of the contrast variation into each subnetwork, converting the scattered field residual into an approximate contrast variation and then enhancing it by a U-Net, thus avoiding the requirement of matched measurement dimension and grid resolution as in existing approaches. The total field and contrast of each layer's output is supervised in the loss function of VBIM-Net, imposing soft physical constraints on the variables in the subnetworks, which benefits the model's performance.In addition, we design a training scheme with extra noise to enhance the model's stability. Extensive numerical results on synthetic and experimental data both verify the inversion quality, generalization ability, and robustness of the proposed VBIM-Net. This work may provide some new inspiration for the design of efficient field-type DL schemes.

Problem

• Full-wave Inverse Scattering
• Accuracy Enhancement
• Robustness Improvement

Innovation

• VBIM-Net
• Variational Born Iteration
• Deep Learning for Inverse Scattering